Intraductal oncocytic papillary neoplasm (IOPN) is a rare precancerous lesion of the pancreas distinct from the more common intraductal papillary mucinous neoplasms (IPMNs). IOPNs always contain high-grade dysplasia and carry significant risk of progressing to invasive pancreatic cancer.
Unlike IPMNs and pancreatic ductal adenocarcinoma (PDAC), IOPNs do not harbor KRAS and GNAS mutations typical of these cancers. Instead they are characterized by fusions in PRKACA and PRKACB genes, now recognized as a molecular hallmark.
Because IOPNs are so rare, their immune microenvironment — the population of immune cells surrounding the tumor — has never been systematically characterized. Understanding this is important for assessing whether immunotherapy could be effective.
The researchers analyzed IOPN cases using whole-slide immunohistochemistry (IHC) staining for key immune markers: CD3, CD4, CD8, CD20, CD163, PD-1, PD-L1, and mismatch repair proteins (MLH1, PMS2, MSH2, MSH6).
An AI-based digital pathology algorithm automatically classified cells in each slide as tumor cells, immune cells, or stromal cells, using smoothed visual features at 25 and 50 micrometer radii with multiple rounds of expert review.
Spatial analysis using Cytomap grouped cells into 200-micrometer neighborhood windows to understand how immune cells are organized relative to each other and to tumor cells — a technique called digital spatial deconvolution.
IOPNs showed substantial immune cell infiltration, with CD8-positive cytotoxic T cells present in significantly higher numbers than CD4-positive helper T cells (mean scores: 4.1 vs. 3.1, p=0.001). This suggests an active cytotoxic immune response within these lesions.
In the two cases with associated invasive carcinoma, the immune landscape changed markedly between the non-invasive (IOPN) and invasive components, with CD4-positive cells reduced and CD8-positive cells increased in invasive areas.
PD-L1, a checkpoint marker targeted by modern immunotherapy drugs, was expressed only in invasive components and at low levels (TPS=5%, CPS=7). All mismatch repair proteins were intact, indicating no microsatellite instability.
The absence of microsatellite instability (MSI) in all IOPN cases is significant, as MSI-high tumors typically respond well to immune checkpoint inhibitor therapy. IOPN patients are unlikely to benefit from this class of immunotherapy based on this criterion.
The low and variable PD-L1 expression suggests anti-PD-1/PD-L1 therapies would likely have limited efficacy in most IOPN patients, though cases with associated invasive carcinoma may warrant individual assessment.
The high CD8 T cell density in non-invasive IOPN components suggests the immune system may be actively suppressing early tumor growth. Strategies to sustain this response during transition to invasion could be therapeutically valuable.
This study is the first systematic characterization of the immune microenvironment in IOPNs, made possible by combining immunohistochemistry with AI-driven digital pathology and spatial analysis. The approach reveals cellular organization that manual counting would miss.
The distinct immune landscape of IOPNs — different from IPMNs and PDAC — reinforces the concept that IOPN is a biologically unique entity requiring its own management strategy.
Future studies with larger IOPN cohorts and prospective follow-up are needed to determine whether immune features here predict progression to invasive cancer or response to specific therapies.